A graphics processing apparatus (often referred to as a GPU—graphics processing unit) may be arranged to perform tile based processing. That is to say, for each frame of graphics data which it generates for display it sub-divides the frame into multiple tiles and performs its graphics processing on each tile. The processed tiles are then accumulated as the frame of graphics data. This is a technique which may for example be used in the context of a data processing system which is intended for implementation in a portable device, because the pressure for such a device to be particularly efficient may be greater. Such tile based processing may require a sequence of smaller memory accesses to be made rather than fewer larger memory accesses, which can be preferable in the context of a portable device. Nevertheless, the pressure for the components of a data processing apparatus to become more efficient in their operations, which may be particularly acute in the context of a data processing apparatus implemented in a portable device, is ever present. The present disclosure provides techniques which can make a tile based graphics processing apparatus more efficient.